The invention provides machine vision systems and methods for morphological transformation of a source image, e.g., with non-uniform offsets. According to these, pixels from selected neighborhoods, or regions, of the source image are rearranged, or unfolded, into respective columns of an intermediate image. For each column in the intermediate image, the method identifies a pixel intensity value of selected rank. Once the pixel intensity value of selected rank is identified for each column, it is retained in a transformation image, e.g., at a location corresponding to center of the respective neighborhood. The foregoing steps are repeated, e.g., along strips or bands, for each neighborhood in the source image.
Legal claims defining the scope of protection, as filed with the USPTO.
1. A machine method for morphological transformation of a source image into a transformation image, the method comprising the steps of A. generating a first intermediate image by arranging pixels from selected neighborhoods of the source image in respective columns of the intermediate image, B. for each of the selected columns of pixels in the intermediate image, determining a value of selected rank as between the pixels in that column by comparing intensities thereof, the value of selected rank being a column winner for that column, and retaining the column winner for that column in the transformation image, C. generating one or more further intermediate images by arranging pixels from further selected neighborhoods of the source image in respective columns of respective such further intermediate images, and D. repeating step (B) for each respective further intermediate image, retaining the column winners determined in step (B) for each respective further intermediate image in the transformation image.
2. A method according to claim 1, wherein step (B) comprises (i) loading the selected columns of a first row of the first intermediate image into corresponding columns in a first data store, (ii) loading the selected columns of a second row of the first intermediate image into corresponding columns in a second data store, (iii) adding a respective offset to each of at least selected columns in the first and second stores, (iv) for each of the selected columns in the first data store, determining the column winner for that column by comparing a value in that column with a value in a corresponding column in the second store, and loading the column winner so determined into a corresponding column of the first data store.
3. A method according to claim 2, comprising for each of the selected additional rows of pixels in the first intermediate image, loading the selected columns of that row into corresponding columns in the second store and repeating steps (B)(iii) and B(iv).
4. A method according to any of claims 1-3, wherein step (B) comprises adding an offset to the intensity of each pixel in a column prior to comparing to it with others in that column.
5. A method according to claim 1, adapted for dilation of the source image, where the selected rank in step (B) is a maximum.
6. A method according to claim 1, adapted for erosion of the source image, where the selected rank in step (B) is a minimum.
7. A method according to claim 1, wherein the neighborhoods are rectangular.
8. A method according to claim 7, wherein the neighborhoods are squares.
9. A method according to claim 8, wherein the neighborhoods are 3.times.3 squares.
10. An article of manufacture comprising a computer usable medium embodying program code for causing a digital data processor to carry out a method for morphological transformation of a source image into a transformation image, the method comprising the steps of A. generating a first intermediate image by arranging pixels from selected neighborhoods of the source image in respective columns of the intermediate image, B. for each of the selected columns of pixels in the intermediate image, determining a value of selected rank as between the pixels in that column by comparing intensities thereof, the value of selected rank being a column winner for that column, and retaining the column winner for that column in the transformation image, C. generating one or more further intermediate images by arranging pixels from further selected neighborhoods of the source image in respective columns of respective such further intermediate images, and D. repeating step (B) for each respective further intermediate image, retaining the column winners determined in step (B) for each respective further intermediate image in the transformation image.
11. An article according to claim 10, wherein steps (A) and (C) comprise generating respective ones of the first and further intermediate images by arranging pixels from selected neighborhoods of the source image in respective columns of the intermediate image.
12. An article according to claim 10, wherein step (B)(i) comprises adding an offset to the value of each pixel in a column prior to comparing to it with others in that column.
13. An article according to claim 10, wherein step (B) comprises (i) loading the selected columns of a first row of the first intermediate image into corresponding columns in a first data store, (ii) loading the selected columns of a second row of the first intermediate image into corresponding columns in a second data store, (iii) adding a respective offset to each of at least selected columns in the first and second stores, (iv) for each of the selected columns in the first data store, determining the column winner for that column by comparing a value in that column with a value in a corresponding column in the second store, and loading the column winner so determined into a corresponding column of the first data.
14. An article according to claim 10, wherein the method is adapted for dilation of the source image, where the selected rank is a maximum.
15. An article according to claim 10, wherein the method is adapted for erosion of the source image, where the selected rank is a minimum.
16. A digital data processing system for morphological transformation of a source image into a transformation image by carrying out a method comprising the steps of A. generating a first intermediate image by arranging pixels from selected neighborhoods of the source image in respective columns of the intermediate image, B. for each of the selected columns of pixels in the intermediate image, determining a value of selected rank as between the pixels in that column by comparing intensities thereof, the value of selected rank being a column winner for that column, and retaining the column winner for that column in the transformation image, C. generating one or more further intermediate images by arranging pixels from further selected neighborhoods of the source image in respective columns of respective such further intermediate images, and D. repeating step (B) for each respective further intermediate image, retaining the column winners determined in step (B) for each respective further intermediate image in the transformation image.
17. A digital data processing system according to claim 16, wherein step (B) comprises (i) loading the selected columns of a first row of the first intermediate image into corresponding columns in a first data store, (ii) loading the selected columns of a second row of the first intermediate image into corresponding columns in a second data store, (iii) adding a respective offset to each of at least selected columns in the first and second stores, (iv) for each of the selected columns in the first data store, determining the column winner for that column by comparing a value in that column with a value in a corresponding column in the second store, and loading the column winner so determined into a corresponding column of the first data store.
18. A digital data processing system according to claim 17, wherein the method comprises the step of, for each of the selected additional rows of pixels in the first intermediate image, loading the selected columns of that row into corresponding columns in the second store and repeating steps (B)(iii)-B(iv).
19. A digital data processing system according to any of claims 16-18, wherein step (B) comprises adding an offset to the intensity of each pixel in a column prior to comparing to it with others in that column.
20. A digital data processing system according to claim 16, adapted for dilation of the source image, where the selected rank in step (B) is a maximum.
21. A digital data processing system according to claim 16, adapted for erosion of the source image, where the selected rank in step (B) is a minimum.
22. A digital data processing system according to claim 16, wherein the neighborhoods are rectangular.
23. A digital data processing system according to claim 22, wherein the neighborhoods are squares.
24. A digital data processing system according to claim 23, wherein the neighborhoods are 3.times.3 squares.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
January 28, 1998
August 28, 2001
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.